Deep Learning-Based Detection of Intracranial Hemorrhages in Postmortem Computed Tomography: Comparative Study of 15 Transfer-Learned Models
Source abstract
Abstract
With the increasing use of postmortem imaging, deep learning (DL)-based automated analysis may assist in the detection of intracranial hemorrhages. However, limited postmortem data complicate model training. This study aims to assess the accuracy of DL models in detecting intracranial hemorrhages in postmortem head computed tomography (CT) scans using transfer learning. A total of 75,000 labeled head CT images from the Radiological Society of North America Intracranial Hemorrhage Detection Challenge serve as the training data for the 15 DL models. Each model is fine-tuned via transfer learning. A total of 134 postmortem cases with hemorrhage status confirmed by autopsy serve as the external test set. Model performance is evaluated using the area under the receiver operating characteristic curve (AUC), sensitivity, specificity, training time, inference time, and number of parameters. Spearman’s rank correlation coefficients are calculated for these metrics. DenseNet201 achieves the highest AUC (0.907), with the AUCs of the 15 models ranging from 0.862 to 0.907. A longer inference time moderately correlates with higher AUC (Spearman’s ρ = 0.586, p = 0.022), whereas the number of parameters is not positively correlated with performance (ρ = −0.472, p = 0.076). The sensitivity and specificity are 0.828 and 0.871, respectively. Transfer learning using a large non-postmortem dataset enables accurate intracranial hemorrhage detection using postmortem CT, potentially reducing the autopsy workload. The results demonstrate that models with fewer parameters often perform comparably to more complex models, emphasizing the need to balance accuracy with computational efficiency.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Autopsy Techniques and Outcomes · Machine Learning in Healthcare · Intracerebral and Subarachnoid Hemorrhage Research
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Shinko Hospital, Kobe 651-0072, Japan
Department of Radiology, Graduate School of Medicine, Kobe University, Kobe 650-0017, Japan
Department of Legal Medicine, Graduate School of Medicine, Kobe University, Kobe 650-0017, Japan
Department of Radiology, Faculty of Medicine, Kindai University, Osaka 589-8511, Japan
Funding
JSPS KAKENHI
References (29)
Show source-provided references
1. Bolliger, S.A., and Thali, M.J. (2015). Imaging and virtual autopsy: Looking back and forward. Philos. Trans. R. Soc. Lond. B Biol. Sci., 370.10.1098/rstb.2014.0253 ↗
2. Post-mortem imaging as an alternative to autopsy in the diagnosis of adult deaths: A validation study10.1016/S0140-6736(11)61483-9 ↗
3. Artificial intelligence in medical imaging of the liver10.3748/wjg.v25.i6.672 ↗
4. Artificial intelligence in cardiovascular imaging: JACC state-of-the-art review10.1016/j.jacc.2018.12.054 ↗
5. Cui, Y., Zhu, J., Duan, Z., Liao, Z., Wang, S., and Liu, W. (2022). Artificial intelligence in spinal imaging: Current status and future directions. Int. J. Environ. Res. Public Health, 19.10.3390/ijerph191811708 ↗
6. Prognosis prediction of patients with malignant pleural mesothelioma using conditional variational autoencoder on 3D PET images and clinical data10.1002/mp.16694 ↗
7. Matsuo, H., Nishio, M., Kanda, T., Kojita, Y., Kono, A.K., Hori, M., Teshima, M., Otsuki, N., Nibu, K.I., and Murakami, T. (2020). Diagnostic accuracy of deep-learning with anomaly detection for a small amount of imbalanced data: Discriminating malignant parotid tumors in MRI. Sci. Rep., 10.10.1038/s41598-020-76389-4 ↗
8. Nishio, M., Noguchi, S., Matsuo, H., and Murakami, T. (2020). Automatic classification between COVID-19 pneumonia, non-COVID-19 pneumonia, and the healthy on chest X-ray image: Combination of data augmentation methods. Sci. Rep., 10.10.1038/s41598-020-74539-2 ↗
9. Kumari, R., Nikki, S., Beg, R., Ranjan, S., Gope, S.K., Mallick, R.R., and Dutta, A. (2020). A review of image detection, recognition and classification with the help of machine learning and artificial intelligence. SSRN J.10.2139/ssrn.3611339 ↗
10. Kim, H.E., Cosa-Linan, A., Santhanam, N., Jannesari, M., Maros, M.E., and Ganslandt, T. (2022). Transfer learning for medical image classification: A literature review. BMC Med. Imaging, 22.10.1186/s12880-022-00793-7 ↗
11. Alzubaidi, L., Al-Amidie, M., Al-Asadi, A., Humaidi, A.J., Al-Shamma, O., Fadhel, M.A., Zhang, J., Santamaría, J., and Duan, Y. (2021). Novel Transfer Learning Approach for Medical Imaging with Limited Labeled Data. Cancers, 13.10.3390/cancers13071590 ↗
12. Li, M., Jiang, Y., Zhang, Y., and Zhu, H. (2023). Medical image analysis using deep learning algorithms. Front. Public Health, 11.10.3389/fpubh.2023.1273253 ↗
13. Virtopsy, a new imaging horizon in forensic pathology: Virtual autopsy by postmortem multislice computed tomography (MSCT) and magnetic resonance imaging (MRI)—A feasibility study10.1520/JFS2002166 ↗
14. Advanced machine learning in action: Identification of intracranial hemorrhage on computed tomography scans of the head with clinical workflow integration10.1038/s41746-017-0015-z ↗
15. A deep learning algorithm for automatic detection and classification of acute intracranial hemorrhages in head CT scans10.1016/j.nicl.2021.102785 ↗
16. RSNA-ASNR 2019 Brain Hemorrhage CT Annotators, Construction of a machine learning dataset through collaboration: The RSNA 2019 brain CT hemorrhage challenge10.1148/ryai.2020190211 ↗
17. (2025, January 26). PyTorch-Image-Models: The Largest Collection of PyTorch Image Encoders/Backbones. Including Train, Eval, Inference, Export Scripts, and Pretrained Weights—ResNet, ResNeXT, EfficientNet, NFNet, Vision Transformer (ViT), MobileNetV4, MobileNet-V3 & V2, RegNet, DPN, CSPNet, Swin Transformer, MaxViT, CoAtNet, ConvNeXt, and More. GitHub. Available online: https://github.com/huggingface/pytorch-image-models.
18. Huang, G., Liu, Z., van der Maaten, L., and Weinberger, K.Q. (2016). Densely connected convolutional networks. arXiv, Available online: http://arxiv.org/abs/1608.06993.10.1109/CVPR.2017.243 ↗
19. Brock, A., De, S., Smith, S.L., and Simonyan, K. (2021). High-performance large-scale image recognition without normalization. arXiv, Available online: http://arxiv.org/abs/2102.06171.
20. Tan, M., and Le, Q.V. (2019). EfficientNet: Rethinking model scaling for convolutional Neural Networks. arXiv, Available online: http://arxiv.org/abs/1905.11946.
21. Tan, M., and Le, Q.V. (2021). EfficientNetV2: Smaller models and faster training. arXiv, Available online: http://arxiv.org/abs/2104.00298.
22. Howard, A., Sandler, M., Chu, G., Chen, L.C., Chen, B., Tan, M., Wang, W., Zhu, Y., Pang, R., and Vasudevan, V. (2019). Searching for MobileNetV3. arXiv, Available online: http://arxiv.org/abs/1905.02244.10.1109/ICCV.2019.00140 ↗
23. He, K., Zhang, X., Ren, S., and Sun, J. (2016). Identity mappings in deep residual networks. arXiv, Available online: http://arxiv.org/abs/1603.05027.10.1007/978-3-319-46493-0_38 ↗
24. Xie, S., Girshick, R., Dollár, P., Tu, Z., and He, K. (2016). Aggregated residual transformations for deep neural networks. arXiv, Available online: http://arxiv.org/abs/1611.05431.10.1109/CVPR.2017.634 ↗
25. Liu, Z., Lin, Y., Cao, Y., Hu, H., Wei, Y., Zhang, Z., Lin, S., and Guo, B. (2021). Swin Transformer: Hierarchical vision Transformer using shifted windows. arXiv, Available online: http://arxiv.org/abs/2103.14030.10.1109/ICCV48922.2021.00986 ↗
26. Simonyan, K., and Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv, Available online: http://arxiv.org/abs/1409.1556.
27. Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterhiner, T., Dehghani, M., Minderer, M., Heigold, G., and Gelly, S. (2020). An image is worth 16 × 16 words: Transformers for image recognition at scale. arXiv, Available online: http://arxiv.org/abs/2010.11929.
28. Cubuk, E.D., Zoph, B., Shlens, J., and Le, Q.V. (2019). RandAugment: Practical automated data augmentation with a reduced search space. arXiv, Available online: http://arxiv.org/abs/1909.13719.10.1109/CVPRW50498.2020.00359 ↗
29. Nishio, M., Koyasu, S., Noguchi, S., Kiguchi, T., Nakatsu, K., Akasaka, T., Yamada, H., and Itoh, K. (2020). Automatic detection of acute ischemic stroke using non-contrast computed tomography and two-stage deep learning model. Comput. Methods Programs Biomed., 196.10.1016/j.cmpb.2020.105711 ↗